Improved Distributed Cooperative Filters for Minimum-Variance Unbiased Estimation of Unknown Inputs and States in Multi-Agent Systems

The estimation of unknown inputs and states is important for multi-agent systems. In this study, we propose an improved filter that overcomes the shortcomings of the existing filter. This filter estimate the unknown inputs in cooperative mode and the states in decentralized mode, respectively. Compared with the former one, the improved filters not only could reduce the amount of information transmitted and calculated, but also have better estimation performance. The optimal gain matrix of the proposed filters is derived and the effectiveness of the filters is shown through simulation.